modules: CMSIS_6: fix path and use it for TF-M and Cortex-M #89370
Conversation
github-actions
bot
added
area: CMSIS-Core
area: TF-M
github-actions
bot
requested review from
d3zd3z,
ithinuel,
microbuilder,
povergoing,
stephanosio,
tomi-font,
valeriosetti and
Vge0rge
|
The following west manifest projects have changed revision in this Pull Request:
Additional metadata changed:
⛔ DNM label due to: 1 project with metadata changes Note: This message is automatically posted and updated by the Manifest GitHub Action. |
github-actions
bot
added
manifest
manifest-cmsis_6
DNM (manifest)
|
Ah. Good idea just enabling it for M only. |
github-actions
bot
added
the
platform: nRF BSIM
wearyzen
force-pushed
the
cmsis_6_update
branch
5 times, most recently
from
7f6707d to
320db7b
Compare
|
The twister failures are observed on main branch as well so they are not related to this PR. |
|
Thanks for this. Could you also put up a PR in the TF-M fork that removes the copy of the CMSIS 6 files (just reverting the commit)? |
|
The last commit to disable the test is just to check if there are any other failures with this PR. However, I'll keep the commit if it is accepted here |
|
Hi @tomi-font , could you please help with review/merge of linked CMSIS and TF-M PRs? |
Sure, will take care of this tomorrow. |
west.yml
Outdated
| - name: cmsis_6 | ||
| repo-path: CMSIS_6 | ||
| revision: 783317a3072554acbac86cca2ff24928cbf98d30 | ||
| path: modules/lib/cmsis_6 | ||
| revision: 6dd50439a9b83398ff2ae1376eef0a2a0b95913b | ||
| path: modules/hal/cmsis_6 | ||
| groups: | ||
| - hal |
There was a problem hiding this comment.
How about moving the cmsis_6 entry right after the cmsis one?
There was a problem hiding this comment.
well, it seems we can't do this as per the compliance check, so I'll have to move them back 😞
wearyzen
force-pushed
the
cmsis_6_update
branch
2 times, most recently
from
3bf8595 to
d7bae60
Compare
What is the change? Fix the path for module CMSIS_6 and use CMSIS_6 module for TF-M. Why do we need this change? After Zephyr updated TF-M to v2.1.0, bb037d4469842c96f5872b271490aceb0734d965 added a local copy of CMSIS_6 to stop Zephyr's TF-M from downloading the CMSIS_6 from upstream. The correct way would be to have CMSIS_6 as a module in Zephyr (which we have now) and pass the path of this module to TF-M. A fork of the upstream CMSIS_6 was added to Zephyr however, the path in west.yml makes it a lib and not a module. Fixing the path generates the ZEPHYR_CMSIS_6_MODULE_DIR symbol which can now be used to pass to TF-M and the copy in TF-M would no longer be required. Signed-off-by: Sudan Landge <[email protected]>
What is the change? Switch to module CMSIS_6 for Cortex-M while continuing to use cmsis module (version 5.9.0) for Cortex-A/R. Why do we need this change? The current cmsis module consists of: - Cortex-M support from upstream cmsis 5.9.0 - Cortex-A/R support which was not upstreamed to CMSIS Upstream cmsis 5.9 was deprecated so we should be using CMSIS_6 however, it seems due to lack of Cortex-A/R support in upstream and other reasons, this was pushed back. While upstreaming Cortex-A/R support to CMSIS_6 could take its time, this shouldn't stop Cortex-M to start using CMSIS_6. Also, if we do not use CMSIS_6 for Cortex-M then using the newer GCC 14.2 toolchain will return below compiler error: ``` zephyrproject/modules/hal/cmsis/CMSIS/Core/Include/core_cm85.h:4406:10: fatal error: pac_armv81.h: No such file or directory 4406 | #include "pac_armv81.h" ``` Using CMSIS_6 for Cortex-M will fix this. Signed-off-by: Sudan Landge <[email protected]>
|
What is the change?
Switch to using CMSIS_6 module for TF-M and Cortex-M while continue to use CMSIS for Cortex-A/R.
Why do we need this change?
Pls refer to the commit messages for more details.
Signed-off-by: Sudan Landge [email protected]